Cambridge Consultants' Deep Tech Index reveals culture and short-termism block innovation, not science
Summary
Richard Traherne and Joe Corrigan of Cambridge Consultants discuss their Deep Tech Index, which surveyed ~750 senior leaders. They identify a 'radical disconnect' where 91% of executives value innovation, yet only 5-10% of R&D budgets fund it, citing bureaucracy and board pressure as primary blockers rather than scientific limitations. The speakers highlight significant sector variance, with semiconductors allocating 36% to radical bets versus 3% in consumer electronics. They argue AI accelerates design but shifts bottlenecks to physical validation and capital allocation. Finally, they predict AI will enable edge compute breakthroughs and shift users from consumers to creators in specialized domains like medical robotics.
Key takeaways
- 1 Sector variance is stark: 36% of semiconductor leaders fund radical bets, compared to only 3% in consumer electronics and 13% in communications.
Unidentified speaker (speaker_1) · 9:19
- 2 Most most people uh name sort of bureaucracy as the thing that that blocks them. And in our report, I think fewer than four percent describe their company as lean and and fast and really able to respond to the innovations that are needed in in time.
Unidentified speaker (speaker_2) · 11:57
- 3 So recently we did some work with um with NVIDIA, with CMR, uh Cambridge Medical Robotics, to to do a well a world's first real-time world model, which which enables robots, surgical robots, to think about um how a surgery is going to play out uh in the future. And that's useful because it it paves the way for new safety cases so that that it can improve the quality of surgery, but it also it it paves the way for autonomy in in surgery.
Unidentified speaker (speaker_2) · 34:20
Episode sections
- 0:18 – 4:22
Introduction and Cambridge Consultants Overview
Host Matthew Bloxham introduces guests Richard Traherne (CEO) and Joe Corrigan (CTO) of Cambridge Consultants, a deep-tech innovation partner now owned by Capgemini. Traherne explains that the firm uses first-principles science and engineering to create commercial products, distinguishing itself from pure consulting firms.
- 4:22 – 8:23
Deep Tech Index Methodology and Findings
Traherne describes the Deep Tech Index, based on interviews with ~750 senior leaders. The study found that while executives are enthusiastic about radical innovation, the primary blockers are culture, bureaucracy, and short-term financial horizons, not a lack of scientific capability.
- 8:23 – 10:04
The Radical Disconnect and Sector Variance
The speakers define the 'radical disconnect' as the gap between high executive ambition for new markets and low actual R&D investment (5-10%). Traherne notes significant sector variance, with semiconductor leaders allocating 36% of budgets to radical bets versus only 3% in consumer electronics.
- 10:04 – 17:10
Organizational Culture and Planning Horizons
Corrigan and Traherne discuss how corporate machinery often inoculates against risk, causing projects to fail due to leadership hesitation rather than technical issues. They highlight 'Erum’s Law' in pharma, where drug approval efficiency halves every nine years despite increased funding, illustrating that structure matters more than tools.
- 17:10 – 24:00
AI as an Accelerator for Deep Tech
Corrigan explains how AI allows Cambridge Consultants to explore vast design spaces, such as engineering biology, accelerating research cycles (e.g., 15 years of work in six months). He argues that while software development costs drop, the bottleneck shifts to physical validation and capital allocation.
- 24:00 – 30:40
Future of Consumer Devices and Form Factors
The discussion turns to AI-driven device innovation. Corrigan critiques early AI devices for merely replicating smartphone functions, arguing that true innovation requires specialized hardware for hyper-personalization, better battery life, and lower latency, moving beyond surface-level interface changes.
- 30:40 – 34:00
Case Studies and Strategic Takeaways
Traherne cites the Lego Smart Brick as an example of deep innovation requiring long gestation and patent protection. He concludes that companies must protect capital from quarterly pressures and appoint 'swashbucklers' to champion and defend radical innovations against corporate inertia.
- 34:00 – 34:20
Future Breakthroughs and Edge Compute
Speaker_2 predicts breakthroughs across a broader range of domains enabled by larger edge compute capabilities. This allows for more complex tasks to be performed at the edge rather than in centralized cloud infrastructure.
- 34:20 – 35:00
NVIDIA and Cambridge Medical Robotics Collaboration
Speaker_2 cites recent work with NVIDIA and Cambridge Medical Robotics to create a world's first real-time world model for surgical robots. This technology enables robots to simulate future surgical outcomes, paving the way for improved safety cases and surgical autonomy.
- 35:00 – 35:59
Shift in Network Effects and User Roles
Speaker_2 argues that AI agents will fundamentally change network effects by shifting users from passive consumers to active creators. Instead of static experiences, the interface will become a micro-specification environment where users generate and share applications and content.
- 35:59 – 37:04
Accelerated Innovation in Medical Devices
Speaker_2 notes that agentic systems will reduce the time to market for surgical innovations. This acceleration will increase the importance of 'surgeon innovators' within the broader innovation ecosystem, a trend expected to apply across all industries.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| Anthropic | private company | neutral | we had finally anthropic filing for their initial public offering, which could come as soon as mid-November. |
| OpenAI | private company | neutral | And also we're still kind of waiting with bated baited breath on what OpenAI is going to. offer up on the back of its |
| NVIDIANVDA | equity | neutral | So recently we did some work with um with NVIDIA, with CMR, uh Cambridge Medical Robotics, to to do a well a world's first real-time world model |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Cambridge Consultants is part of the Capgemini group.
Unidentified speaker (speaker_1) ·3:15
How to verify: Check Capgemini's investor relations page or recent press releases for acquisitions/partnerships with Cambridge Consultants. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Capgemini has approximately 420,000 people around the world in over fifty countries.
Unidentified speaker (speaker_1) ·3:15
How to verify: Verify against Capgemini's latest annual report or corporate website statistics. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
The Deep Tech Index interviewed about 750 senior leaders across innovation, with 90% being C-Suite.
Unidentified speaker (speaker_1) ·5:30
How to verify: Review the published 'Deep Tech Index' report by Cambridge Consultants for methodology and sample size details. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
No one in the Deep Tech Index survey said that science was the big blocker to innovation.
Unidentified speaker (speaker_1) ·6:10
How to verify: Review the published 'Deep Tech Index' report for the specific breakdown of blockers cited by respondents. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Exploratory work receives only about five to ten percent of R&D budgets for the surveyed CXOs.
Unidentified speaker (speaker_1) ·8:02
How to verify: Review the published 'Deep Tech Index' report for budget allocation statistics. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
36% of semiconductor leaders put most of their budget into radical bets, compared to 3% in consumer and 13% in communications.
Unidentified speaker (speaker_1) ·9:19
How to verify: Review the published 'Deep Tech Index' report for sector-specific budget allocation data. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Fewer than four percent of companies describe their company as lean and fast.
Unidentified speaker (speaker_2) ·11:57
How to verify: Review the published 'Deep Tech Index' report for organizational readiness scores. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
92% of interviewees named board and investor pressure as a key challenge.
Unidentified speaker (speaker_1) ·12:37
How to verify: Review the published 'Deep Tech Index' report for challenges to innovation statistics. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
NVIDIA and Cambridge Medical Robotics collaborated on a world's first real-time world model for surgical robots.
Unidentified speaker (speaker_2) ·34:20
How to verify: Search for press releases or technical publications from NVIDIA or Cambridge Medical Robotics dated between 2026-08-20 and 2026-11-20 mentioning 'real-time world model' and 'surgical robots'. Check NVIDIA's newsroom or Cambridge Medical Robotics' official website. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Primary sourceThe real-time world model enables surgical robots to predict how a surgery will play out in the future.
Unidentified speaker (speaker_2) ·34:20
How to verify: Review technical documentation or white papers from Cambridge Medical Robotics or NVIDIA detailing the capabilities of their surgical AI systems. Verify if the system includes predictive simulation features. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
The technology paves the way for new safety cases and autonomy in surgery.
Unidentified speaker (speaker_2) ·34:40
How to verify: Check regulatory filings or safety certifications from medical device regulators (e.g., FDA, EMA) or company press releases regarding 'autonomy' or 'safety cases' for surgical robots. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Primary sourceAI agents will shift users from passive consumers to creators within network effects.
Unidentified speaker (speaker_2) ·35:19
How to verify: This is a forward-looking assertion. Verify by tracking industry reports or product launches from major tech platforms that introduce 'creator' tools driven by AI agents during the verification window. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Do your own research
- Review the full Cambridge Consultants Deep Tech Index report to verify the methodology and sample size of the ~750 senior leaders surveyed.
- Check quarterly earnings reports of major semiconductor vs. consumer electronics companies to validate the claimed R&D allocation disparities (36% vs 3%).
- Verify the partnership between Cambridge Consultants, NVIDIA, and Cambridge Medical Robotics through official press releases or press kits.
- Research 'Erum’s Law' in pharmaceuticals by reviewing FDA approval timelines and R&D spending data over the last few decades.
- Investigate the 'Whoop' fitness tracker's origin story and patent filings to confirm the Cambridge Consultants prototype history.
- Analyze the financial performance of companies that have appointed dedicated 'innovation champions' versus those that have not, to assess the 'swashbuckler' hypothesis.
- Look for counter-arguments to the claim that bureaucracy is the primary blocker by reviewing case studies of failed deep-tech startups due to technical, not cultural, reasons.
Full transcriptexpand
0:18Unidentified speaker (speaker_0): Hello, and welcome to the Bloomberg Intelligence Tech Disruptors Podcast. My name is Matthew Bloxham, I'm a senior analyst. covering technology and media. And for those who are not familiar with us, Bloomberg Intelligence is the investment research arm of Bloomberg with approximately 500 analysts and strategists covering all major global markets and asset classes. I'm delighted to be joined today by two executives from Cambridge Consultants. We have Richard Trahan, Chief Executive Officer. and Joe Corrigan, Chief Technology Officer. Richard, Joe, welcome to the podcast. It's great to have you with us today.
0:58Unidentified speaker (UNKNOWN): Well,
0:59Unidentified speaker (speaker_1): Matt, thanks for the invitation, and we're we're really pleased to be on. Great.
1:02Unidentified speaker (speaker_0): Well, look, I'm really excited for this conversation. But before we get into the detail, I think it'd be really helpful if you could uh give us a little bit of background on Cambridge Consultants, what the company. does. Uh, I know in the last few years the ownership uh structures changing, you're now part of the CAP Gemini group. So, yeah, that'll be that would be, I think, a helpful introduction for for our listeners.
1:25Unidentified speaker (speaker_1): Yeah, sure. So uh Cambridge Consultants, or or CC as we call ourselves, um, essentially helps its clients to dramatically increase their value and their leadership through world-changing breakthroughs. And we do that by taking business ambitions and moving them from sort of first principle theory and physics and science through the prototype stages all the way through to products and services that are a commercial success in the real world. And I emphasize the commercial, it it's not inventing for the sake of it, it's creating commercial outcomes. And we we work with a range of bold rebel type leaders. We we turn them swashbucklers within the business who have a real ambition and drive to change the status quo. And importantly, they see a deep tech strategy, which is what our business is centered on, as a winning choice of all of the choices they could have to reinvent their market. And so we sometimes say that through that very first principle approach, we're making
2:27the seemingly impossible possible. And of course, we're doing that today in the AI era. And later, I know Joe will talk a lot more about that. But AI is both a technology that we use in the solutions that we create for clients, but it's also a tool that we're using to disrupt how we work
2:49Unidentified speaker (UNKNOWN): and
2:49Unidentified speaker (speaker_1): indeed how our clients are working. to develop uh new RD and other forms of innovation.
2:55Unidentified speaker (UNKNOWN): Um,
2:55Unidentified speaker (speaker_1): so in our name, you you know it's got the word consultants, and that's often uh misunderstood people picture advice, but we're an innovation partner that's built on science, engineering, design, labs uh that create that that outcome. For us, that's really important distinction because today, if you're uh as you well know, if your product is a clever a person, person's time, AI is likely to eat your business. And that's not our business. Uh you mentioned Cap Gemini, uh, so our parentage and Cap Gemini is uh an AI powered business and technology transformation partner, uh around 420,000 people around the world in over fifty countries. And so that's uh a a great parent to have to bring scale and a level of digital transformation that is sort of industry changing to our clients with global reach. Great.
3:47Unidentified speaker (speaker_0): Well, that's a fantastic introduction. I'm just I'm going to bring in Joe briefly there. I mean it being the chief technology officer at an incredibly technology heavy business must be almost like being in a toy shop, maybe, is it?
3:59Unidentified speaker (speaker_2): It is a tremendous place to work at. Yeah. There's a there's a huge number of brilliant engineers and scientists here. We've got uh about seven hundred and fifty people, most of them scientists, technologists who are working on breakthroughs for our clients and also on our own research. Um so we're really trying to push state of the art in a number of key domains. So yes, it's a it's a tremendously exciting place
4:21Unidentified speaker (UNKNOWN): to
4:21Unidentified speaker (speaker_2): work. Yeah.
4:22Unidentified speaker (speaker_0): Well, let's kind of get into the the meat of our conversation today. And you know, One of the reasons that we set this conversation up was to talk about a report that you published recently, the Deep Tech Index, which I think was published in mid September. Maybe as a starting point you could uh give us a a top level description of what that report is about, you know, the motivation behind commissioning it and perhaps, you know, one or two of the kind of more interesting findings that that you you got from the research.
4:50Unidentified speaker (speaker_1): Yeah. sure. So the Deep Tech Index was basically built on the premise that we've you know long understood the benefits of using first principle innovation. You know, in our history, we've helped to create the industrial inkjet printing era. You know, look at any printing on a water bottle or a food packet. You know, that that's come from our original work. Um, the Bluetooth in your in your pocket, in your phone, is very likely derived from Bluetooth that we helped to create through Cambridge Silicon Radio, a spin-out of of CC. So, you know, these major big foundational steps that we've taken from first principle science. And about two years ago, we asked ourselves, you know, why aren't more companies using radical innovation as a business strategy? And so we uh set up the index to look for that answer. We interviewed about 750 senior leaders across innovation. Uh around 90% of those were C Suite. And of those, we had 25 in depth interviews as well. And we fundamentally asked one question you know, what's stopping your company doing anything genuinely new? And we had a lot of feedback. Uh, we coded 973 separate answers. And the interesting thing was that not one
6:10person said that science was the big blocker. They talked about culture, they talked about red tape, and they talked about the clock, you know, our proxy for timescale. scale. And so, in other words, you know, a lot of this was was to do with the company's attitude to risk. And so, you know, Joe, you might want to talk a little bit more about how that sort of feeds into to our work.
6:31Unidentified speaker (speaker_2): Yeah, so that's why we we framed it around um deep tech and radical innovation. So deep tech starts from a really hard scientific or engineering problem. And it takes years of sustained risk to get it working. And from that it's very, very hard to copy.
6:50Unidentified speaker (speaker_1): Yeah. And I think that just to underline the, you know, the long time, you know, more than a decade ago, three founders came to us. They were sitting on Boston, uh, east coast of the US, and they wanted to measure continuous heart rate from the wrist, um, accurate during hard exercise. They were targeting elite athletes with no chest strap. And you know, you've got this problem of trying to measure that tiny signal in skin colour. changes. Uh the m movement that you're doing as exercise buries it in noise. And fundamentally, it wasn't clear how that could be done. Anyway, our our team helped them to do it. That first prototype proved it could be done, and that prototype became whoop. Uh and they've now got to their uh major valuation. Uh and but that's been a decade of work, and we can point to other clients where you know that that that timescale timescale has has been important. And so going back to the index and relating it to that, um, nearly six in ten of the CXOs told us that radical innovation is is the most exciting kind of information uh innovation for them. You know, their belief innovation was high, but uh I think it was 91% Sense said it it mattered to them personally. But in our interviews
8:01Unidentified speaker (UNKNOWN): with
8:02Unidentified speaker (speaker_1): the with the that cohort, the exploratory work in in that area was only about five to ten percent of the RD budgets that they were uh in charge of. And so that's the gap. You know, that that gap between the ambition and the level of investment and progress is is what we were studying, you know, for the past year with the report.
8:23Unidentified speaker (speaker_0): Yeah, and you talk about that gap, and you the radical came up quite a few times. in the report and you've mentioned kind of radical in innovation there, but also this idea of radical disconnect. And i is that kind of where you're talking about this kind of gap between the ambition and maybe the kind of ability to kind of deliver on that that goal?
8:45Unidentified speaker (speaker_1): Absolutely. So yeah,
8:47Unidentified speaker (UNKNOWN): you
8:47Unidentified speaker (speaker_1): know, that radical innovation is typically creating not just a new capability, but a a new market. You know, it's where organizations are looking to go beyond what they do every day and to create an entirely new business opportunity. And that's, you know, driven by of course combining new um and and existing technologies in ways that haven't been done. before. But it's also about the journey to get that from the idea boardroom out to the market. And that as as exactly as you're saying, Matt, that that's the gap that we call the radical disconnect. And it varies hugely by sector. You know, i in the report, uh, I think it was 36% of semiconductor leaders put most of their budget into radical bets, but in consumer, it was a very low. three percent. And consumer uh sorry, communications was also quite low at about thirteen percent.
9:45Unidentified speaker (speaker_0): And um looking through the report, you know, th uh there was a great summary towards the back, this kind of overall kind of readiness score with six different factors leadership organizational, financial, and boardroom, boredom, cultural index, uh structure and processes, and an external and technical index. Is there do you have like um a consensus or a strong sense of you know which of those different levers either the the easiest to kind of make progress on or would maybe have the kind of biggest impact in moving people towards successful radical innovation?
10:19Unidentified speaker (speaker_1): Yes. I mean it
10:20Unidentified speaker (UNKNOWN): it
10:21Unidentified speaker (speaker_1): is interesting because if we sort of integrate those results together, the top theme was really the sense that innovation champions, you know, th those swashbucklers who were prepared to champion something that was different, you know, challenge the status quo, were reporting that they were watching their leadership, you know, the people around them playing visibly safe. So it wasn't that they they were personally afraid of of failure. or or career risk or or anything like that. They were kind of reading that corporate room and that room was was saying no.
10:57Unidentified speaker (speaker_2): Yeah and and this is a problem that that compounds so and it's it's one that we've seen um with the clients that we work with is if you if you squash one really radical idea, one really good idea, and the company kind of remembers that and is less bold next time. And so you have the will to do this amazing thing, but the machinery of the the corporation is inoculated against taking these big risks. And so on the projects we deliver, they rarely fail on on the technology. They rarely rarely fail because um it the integration of technology fails. Sometimes they fail because the physics just can't be made to work, but that's that's pretty rare. What really makes them go are the the leaders who are this kind of unstoppable force and they can carry it through all the way to market. Where we fit with that is if if you're a senior leader that finds themselves blocked by the machinery. So we've built these teams and the process that's constantly rewarding that kind of that kind of innovation. And it can help leaders build from the the the reason to believe all the way through to that product launch. Most most people uh name sort of bureaucracy as the thing that that blocks them. And in our report, I think fewer than four percent describe their company as lean and and fast and really able to respond to the innovations that are needed in in time.
12:17Unidentified speaker (speaker_1): Yeah. And I think that's a a theme that we see running through you know our day-to-day work and and have done you know for a long time. And and then another theme maybe to bring into focus is the clock. So every one of our 25 interviewees raised the if you like the two to three year planning horizon. And we had a really startling statistic that you know 92% named board and investor pressure as as a key challenge, Matt, as you as you're highlighting in in terms of you know what are those what are those blockers. So you've kind of got a company that believes in the long term, but it's funding on a 30 month leash. And in our business, you know, we we operate from you know healthcare through telecoms, through consumer, you know, energy and industrial, but but one sector where you know there is a great example is pharma. So in the pharmaceutical sector, you know, that's really shown how that can play out because you know from the 1950s, nineteen fifties, you know, the tools have got dramatically better. You know, it's a market sector that's, you know, extremely well funded. Um, and yet the drugs approved per billion dollars of R and D have roughly halved every nine years. And so, uh, you know, that's been dubbed Erum's law. E Erum being Moore's law spelt backwards, uh, yeah, trying to emphasize that you've got this halving rather than
13:37the, you know, the increasing positive rate that that Moore's law applied to semiconductors. So so there you've got an industry where you've had you know 60 years of of near unlimited money, uh the tools weren't the constraint, um uh, you know, and the structure around them was,
13:53Unidentified speaker (UNKNOWN): um,
13:53Unidentified speaker (speaker_1): you know, which really demonstrates that you you know you can't um cost cut your way to something that doesn't yet exist if you're looking to create a breakthrough.
14:01Unidentified speaker (UNKNOWN): Yeah,
14:01Unidentified speaker (speaker_0): that's fascinating. I've not heard of that. Aroom's law before. I'm definitely definitely taking that one away. And I can definitely vouch from my own experience from all the companies I've worked for in my career that that funding capacity and also public versus private definitely has a big, big impact on the way a company thinks and behaves, particularly when it comes to innovation and investment. Before we move on to the next section, actually I just wanted to kind of touch on something actually, uh Richard, you mentioned you know some differences there between different sectors. Were there any particular differences you saw between either regions or the size of company in their approach to innovation?
14:42Unidentified speaker (speaker_1): Yeah, it's a very pertinent question because let let me deal with size. With with size, uh you know, clearly the strength of a large organization is its scale and breadth. And equally, I think we'd all recognise that that comes with a certain amount of structure and process. And you know, when I talked earlier about that corporate room, that corporate room becomes a a stronger element. And so we work specific, speci you know, when we're working with organizations around the world, we're typically doing two things. You know, we're yes providing science, technology, innovation, design, all of those ingredients to produce a breakthrough, but we're also working with that organization to help them effectively buy that to implement it and you know ha ha have nurture nurtured it and grow it within the organization. um And also for typically the the swashbuckler uh rebellious uh people who are prepared to change the status quo to work with them to help the rest of the organization who may not be of that profile uh to do that. So that
15:51that's one one uh uh aspect. Um you mentioned region and there you know something I think we we we we love working with different cultures around the world because I I won't uh put names on them, but uh you know that there are certain some cultures around the world which like to start quickly, run fast, and get to something. And you know, that is a great way to get an innovation off the ground to you know to prevent that stasis that uh can have everybody sort of Or staring at their navels and and wondering whether it's a good a good idea. Equally, there are you know there are cultures. around the world, and I would put the UK in this category where there's a lot more upfront uh specification, thinking, feasibility work, and things like that.
16:40Unidentified speaker (UNKNOWN): And
16:40Unidentified speaker (speaker_1): so one of the other things we look for is to create a mix. And uh yeah, that's worked really well in in in certain situations where we've been able to blend those different attributes together. Uh and then within, you know, click clearly startups are a different category. We work with many startups to help them take their core idea and you know get that to uh commercial success. Uh and there, of course, the dynamic is very different. You know, there it's uh there's the often, you know, lots of ambition, but it's more about the infrastructure to, you know, to put around it.
17:10Unidentified speaker (UNKNOWN): Got
17:11Unidentified speaker (speaker_0): it. Thanks for those insights. That's kind of fascinating. Now I want to move on to the next section of our conversation where we're going to bring in AI. It's impossible to go a day without kind of um reading at least a few things on AI. And actually, you know, that this week and last week been really in the spotlight, partly because of uh all of the r reporting, but also because we had finally anthropic filing for their initial public offering, which could come as soon as mid-November. And I was reading through that prospectus. I think it's about 300 pages long. So there's a lot to digest. Probably about a third of it is risk factors. But one thing that kind of really stood out to me was some comments that they made that yeah, AI is obviously a massive innovation in its own right, but that actually AI has a really important role to play and as they described it, an innovation in the method of innovation. And actually that kind of connected with something that I read in the Deep Tech Index, which was this idea of AI native R and D. So maybe this is a great, great time to kind of bring Joe back into the conversation and maybe kind of um, you know, kind of explore that a bit, you know. how you think about AI kind of
18:29transforming the approach to innovation and how you see AI native R D.
18:37Unidentified speaker (speaker_2): I mean, it's a it's a tremendously exciting time at the moment because the tools that we have from Anthropic and from from others, they complement the kind of AI tools that we've been developing for the last twenty 20 years at CC. So we we our AI team is is long established in developing machine learning as it used to be called and AI today. And those tools in combination with our own tools mean that we can address new types of data, we can do we can do new types of research, and we can do it at a speed that is that's that's unprecedented. And it means that we can do more experiments. And so really the advantage that you get is particularly in deep tech is is this ability to um explore a search space. And so what we've been doing so for as an example of that is in engineering biology there there's there's more um DNA sequences uh th than atoms in the observable universe and scientists have been sampling that space uh for years and years. Um and using their intuition about uh intuition about how to how the molecules go together uh and and validating with that with work at the bench. And so the AI just lets us search more of that space and it helps us also to um to generate and rank candidates in a way that we haven't
19:56been able to do before. So it creates what we call high leverage points within our data, so we can deliberately design experiments that feed. an AI, knowing that we've got an AI downstream of our experiment. So in our own lab, we've been working on um plastic degrading um enzymes. So we built a protein foundation model and then we uh pointed the AI agents um at our lab and we developed the interfaces between our our physical lab that we have here in Cambridge and and our AI, and we fully automated that process in order to to to iterate that research. And we did about 15 years worth of research in about six months. It's a massive acceleration to the type of work that we can do. And in fact that that work was featured by Microsoft in its recent build uh conference. And the other thing to say is that most companies have used AI so far to make sort of individuals faster. But if everybody's got access to the same model, it doesn't just it doesn't really lead to a competitive advantage. So over the last two years, you'll have seen more VC money going into deep tech than the whole of the previous decade, because essentially the development of of software based um technology has become close to free. The the the software development cost has has decreased massively. And we see that reflected in the survey again. So every
21:16investment in in our survey below AI is physical. So it's either hard science or or deep tech or engineering in in some way. The thing that AI really does for us is it um it does allow us to to do more experiments. It allows us to see further, but but humans really still have to point the telescope. scope.
21:33Unidentified speaker (speaker_0): Yeah that's an interesting point. And Richard what what what are your perspectives on the the role of AI and shaking up, I guess, the the the way companies innovate?
21:43Unidentified speaker (speaker_1): Yeah, well I uh really echoing you know what Joe said, I think it it's fundamental because you know that that aspect of the design space, you know, if if you're looking to create the best innovation in the world, you know, you want to enjoy the largest opportunity space and and and then design space to to work in. And as Joe was saying, you
22:04Unidentified speaker (UNKNOWN): know,
22:04Unidentified speaker (speaker_1): if if that exploration using AI, you know, using AI to look at more atoms than in the observable universe, um, in biology, for example, if you're using it in that way, you know, what what what becomes scarce? You know, what what then is the differentiation? You know, and and also what's the challenge?
22:24Unidentified speaker (UNKNOWN): You know,
22:24Unidentified speaker (speaker_1): if we're all equipped with with that form of AI. what what becomes the you know the challenge that can set people apart? So you know we we very much have this feeling that the bottleneck is shifting at the moment and it it moves to deciding what questions to ask of the design space, uh putting more upfront effort into the concept and you know business, uh not not just technical, but but business concepting. Um, and then critically to proving those answers in a physical reality. Because you know, if you think about a lot of the applications that we read about and see on the television and so on, a lot of those are based in a digital reality and where they become physical in robotics at this point in time, you know, it it's all relatively basic. Um, so you know, bringing these into a physical reality and taking that next step, and then fundamentally putting significant capital and time behind them, you know, that that's really the pharma lesson again. Uh, you know, the
23:24tools didn't fix it. Um, it was more the infrastructure around them. So, so we do think that AI could Would widen the gap in the index. You know, this gap between ambition and you know. commercial outcome. And companies that will win in that environment are those that can fund and and then protect that radical work uh generated by AI and they'll pull ahead faster.
23:53Unidentified speaker (speaker_0): I Richard, you mentioned earlier in the conversation, which has you know been a fantastic innovation for I well, not just profess professional athletes, but uh amateurs too, who uh take their um sport seriously. And you know, it's uh been an innovation in a consumer device. And I guess one of the big things that we're always looking at, and quite often get asked, is you know, what what's gonna is AI going to kind of really uh shake up the form factor of everyday devices we use, both in our personal business life. lives. And obviously, the kind of the big one that people always ask about is the smartphone. And, you know, I guess in the last year or so we've seen some attempts at innovating new form factors. I guess the most notable recent one is Meta's Muse Charm, which is linked to Muse, which is their new agentic Authentic AI assistant. And also we're still kind of waiting with bated baited breath on what OpenAI is going to. offer up on the back of its collaboration with Johnny I from Apple. I think that we have got some early tease of a donut-shaped kind of device. And in the thick of all that, you know, Apple's latest innovation is um a foldable smartphone to which so it feels like you know the the smartphone's still going to be incredibly important. But you know, g given your
25:12deep involvement as uh Cambridge consultants with companies thinking about these challenges, how do you see uh the opportunity for AI to kind of really shake shape up these devices, maybe whether it's general compute or more kind of application specific? Um, and kind of do you see us waiting, you know, many years before there's a real breakthrough, or do you think the pace of change is gonna going to accelerate?
25:37Unidentified speaker (speaker_1): Yeah, well I'll bring Joe. in in in a second because this is very much in his area. But I I think Matt, you're highlighting a a really fascinating area because it is one of these areas that's a paradigm shift. You know, we have the opportunity with AI to you know make step changes in the way that we approach uh you know that Yeah, human technology interaction. And I I I like the fact that you you kind of ended the question in in both sequencing and in the platform architectures that will practically uh enable that uh over time. But may maybe Joe, yeah, this is very much in your area. Do you want to jump in?
26:16Unidentified speaker (speaker_2): Uh yeah. I mean, again, it's um it's very hard to predict the future in in this space, but we we have been working on some some new products in this space and it and it really transforms. the way that you think about doing design. So we've we've talked about this this notion of hyper personalization. You can have a device that really works for you. It's an agent working on your side, but what we've seen so far has been a little bit disappointing. Some of the early versions of, you know w I'm not talking about Meta's Muse necessarily. That that that's yet to to have its date. day, but um some of the previous iterations of this kind of thing, they essentially tried to achieve the same job as a mobile phone. They packed in all of the same kind of technology. And that led to a subpar user experience because you're trying to solve the same the same problem but e using a new interface. It doesn't go far enough. Uh and what Meta's doing is is interesting because it's it pushes you to have that agentic agenic support beside you. But I think you can go even further than that. And we're we're looking at how we do that. And the other thing to say about that is if you focus, if you if you think if you rethink what you're trying to do with the technology and you focus really on the person, it means that you can you can specialise the hardware in a particular way. And that means that that you can get better battery life, you can get
27:36better privacy, you can get lower lower heat, uh, for example, you can get much, much lower latency. Um, and then you can increase privacy and you can inc increase trust. So the reason I think that the previous devices sort of failed in this space is that they didn't really they didn't really focus on try to break through. They focused on on kind of just the interface. It was a it was a it was a surface level innovation. But if you can really if you can really push the limits of what can be done with the physics of how much compute you can get at the edge how much compute is available and how much intelligence is available to you via this via a hybrid system then you can get to a system where yes the interface is specialized but it's a general purpose device that can be tuned to meet this sort of long tail of micro applications this uh ultra personalized personalised need that everybody has to to get particular jobs that are unique to them done in their everyday life.
28:32Unidentified speaker (speaker_0): Um this might well I'm gonna I'm gonna ask this question. I'm kind of not feeling too hopeful about about the response, but I mean are you working on pro projects right now that you kind of feel, you know, really excited that, you know, maybe I don't know on a year or two that that everything's gonna come together and that is really gonna kind of be a radical innovation in the devices that we either use in work or in our personal lives.
28:58Unidentified speaker (speaker_2): Um I can't talk to I can't talk to devices that that we're actually working on. But I can tell you about the the general trends in the space. So the the speed of compute and the speed of compute particularly at the edge means that it's likely that we're going to get devices that can do much, much more than we've ever seen. For also if you look at the general trend in energetic capability over the long term, so the the meta analysis, so this is M E T R analysis, shows a a doubling uh of performance every seven months or by four months if you take a slightly different view. But so that is an incredible pace of change in
29:37Unidentified speaker (UNKNOWN): terms
29:37Unidentified speaker (speaker_2): of the the time horizon of task completion that a model can complete. And if they can do that, what that means is that you can you can give a much more complex task to the system. So uh in our labs, we have a agenc teams of of 25 or more um agents, all set up to take different parts of a program, break it up, and then develop a solution, particularly for a particular area. end user. And so that completely transforms what that end user can achieve. So it's a it's a very, very different experience to being served a kind of one size fits all user interface. It's you you get the tailored experience every single time and you can get you can achieve things that you could never achieve before.
30:20Unidentified speaker (speaker_0): Well I'm excited to kind of see some of these innovations event eventually come through uh to the market and hope hopeful hopefully you know, before too long we'll be able to have another conversation where you can actually talk about some of these products that have um made it into market. We are kind of coming up towards the close of our conversation, but but before we do close up, I just kind of wanted to kind of ask a couple of additional things. I mean, first up, we're just like a really an open invitation to to bring in, you know, something from the report or just from uh your work as Cambridge consultants that you think would be interesting and relevant to the conversation we've had. And I guess the other question was, you know, thinking ahead of year, I'm assuming that that you may well run the uh the index again. You know, if you had to kind of get your crystal ball out and think about, you know, where most progress had been made against uh the index, you know, what what where would that have been?
31:15Unidentified speaker (speaker_1): Sure. So well, maybe to throw in another practical example, uh, earlier this year
31:22Unidentified speaker (UNKNOWN): the
31:22Unidentified speaker (speaker_1): Lego Group announced its new smart play system, which included elements like the smart brick, and that was uh a platform that uh we supported them in. Now let the Lego Group is is an outstanding, amazing brand. And to be part of that was very special. But in that smart brick, it holds 25 patented world firsts. It's the biggest change to the Lego system in place since the minifigure in 1978. So that's the level of innovation that can happen uh again over a long gestation period uh, when you bring first prints. principle science and technology together and implement it in in in such a a simple way that uh well both both kids and adults can can enjoy. I mean in terms of the the report, I think you know that the big takeaways for us were really those elements of protecting
32:23Unidentified speaker (UNKNOWN): a
32:24Unidentified speaker (speaker_1): slice of capital from that quarterly driven clock, you know, just finding the amounts are going to survive a bad quarter, as some of our clients often say. Um, you know, putting a swashbuckler or, you know, that that kind of profile behind it, someone who's really going to challenge the status quo and and critically defend that work in public,
32:43Unidentified speaker (UNKNOWN): you
32:44Unidentified speaker (speaker_1): know, to not allow it to become that initiative that slowly dies in in the corner of the corpus. corporate room. And then get going out of your way to you know to make sure it's defendable because
32:57Unidentified speaker (UNKNOWN): you
32:57Unidentified speaker (speaker_1): know too often we see you know apparently you know great new breakthrough ideas. You know, take a an Uber for s for example, but but you know then you have a lift or a grab you know following what what can you do to ensure that you're injecting a defendable protectable position into what you're doing? And that's something that it is a core part of what we do with all of our clients because unless we give them that, we know that the value of what we deliver, you know, will erode quickly over time. So, you know, back to the report as uh as you were flagging, you know, that that's the gap Yeah, that we're talking about. You know, that sort of six in ten that are excited and and fewer than one in six that that fund it, over time we see AI accentuating that gap potentially, because things are going to become tougher and more complex and and that's where you really have to think about, you know, the the the commercial outcome and and drive everything towards that. Joe, what what what would you add?
34:00Unidentified speaker (speaker_2): So the thing we were talking about the the breakthroughs that we'll see coming in the future. And there were a couple of things that I wanted to touch on that I didn't touch on previously. One of them is that we're going to see we're going to see breakthroughs in a broader range of domains. So as we get larger edge compute, it means that we can do we can do bigger and better tasks at the edge. So recently we did some work with um with NVIDIA, with CMR, uh Cambridge Medical Robotics, to to do a well a world's first real-time world model, which which enables robots, surgical robots, to think about um how a surgery is going to play out uh in the future. And that's useful because it it paves the way for new safety cases so that that it can improve the quality of surgery, but it also it it paves the way for autonomy in in surgery. And the second the second part. that I wanted to touch on was that we we were talking about defensibility or Richard was talking about defensibility just just then. And one of the one of the big drivers particularly in the tech space is that sense of a is is a network effect around around a particular product or around a particular service. So this is the interactivity between users of that service. And because we are experiencing this shift where agents can take more of that
35:19workload, it does completely change the interfaces. So what we're likely to see is instead of a system whereby you're served a particular use case and then you interact on that network in a particular way, people become creators within that space. So instead of it being a static experience, it's a micro-specification. much more user-generated content style experience. So you can imagine people developing their own applications and sharing their own applications with each other. So going beyond, you know, gifts and the kind of fun, the fun interactivity that we've seen today, but really into sharing insightful content. So we've we've long had this concept in surgery of the surgeon innovator, and you have these you have these particular innovators who will will push the boundaries of of what can be done. And it takes a long time to get a surgical innovation out to market. But now with an agentic system, actually that that time to market is going to be reduced and those surgeon innovators are going to become more and more important in the innovation ecosystem. So that's just one small slice through it, but we'll see that kind of story playing out across across every industry that we're in.
36:27Unidentified speaker (speaker_0): That's uh a very kind of upbeat, optimistic um way to kind of think um close out our our conversation. Definitely looking forward to an acceleration of exciting innovations that yeah are positive contributions to the human existence in in in the coming years. Um so Richard, Joe, thank you very, very much for joining us today on this episode of uh Tech Disruptors. And uh yeah, hopefully we can uh do it again sometime next year to catch up on how the business is going and and where we are in the world of AI and innovation. So thank you very much.
37:00Unidentified speaker (speaker_2): Thanks very much. Looking forward to it.
37:02Unidentified speaker (UNKNOWN): Well,
37:02Unidentified speaker (speaker_1): thank you, Matt.