Fueling the Future: Why AI Partnerships are Critical for UK Startups
We all know that the emergence of gen-AI marks a fundamental shift, it will transform economies and how we all live and work. However, the immense resources required: compute power, vast datasets, and specialised talent are often beyond the reach of individual startups.
Today, Startup Coalition is publishing a new paper, Fuelling the Future: The Role of AI Partnerships, which examines the role these partnerships play in bridging this gap to accelerate innovation. And, how the Competition and Markets Authority (CMA) should respond.
For AI startups, collaboration with established players is a lifeline, offering benefits that even traditional Venture Capital (VC) often cannot provide quickly enough. These partnerships provide three critical elements:
Enhanced Financial Capacity: Partnerships act as a key capital injection for sustained R&D and attracting top-tier talent, filling a role often sought from VC funding.
Unfettered Access to Computational Infrastructure: A significant benefit is access to powerful compute from “hyperscalers” like Alphabet, Amazon, and Microsoft. Structured deals substantially lower the cost of training and developing AI models, democratising access and fostering a more competitive market by lowering the barriers to entry.
Accelerated Market Entry and Model Refinement: Partnerships offer a crucial testing ground to trial and improve models, enabling startups to rapidly deploy solutions, gather real-world feedback, and iterate quickly—all essential for creating practical AI solutions.
The complex and interdependent nature of the AI value chain, from infrastructure to the application layer, makes these collaborations crucial. The UK is expected to accrue most value in the application layer, which relies heavily on strong foundations and the ability to fine-tune models.
These partnerships have drawn close scrutiny from the CMA and competition authorities across the world. It is right that they look closely at how these partnerships are struck, structured and how they develop. However, this examination should be done at an arms-length, without impeding or sending unnecessary flares up that may chill a nascent ecosystem.
The CMA has dedicated significant effort to understanding these new dynamics, which it views through the lens of potential competition concerns. In our report we offer some pseudo-psychology on that – after all many have critiqued what have been viewed as ‘past regulatory lags’ in digital markets. However, balance is key here lest the pendulum swings far too far in the opposite direction.
We are seeing signs that this is happening and the risk is it puts the brakes on just at the time we should be hitting the gas. Prolonged investigations, or even the threat of them, force agile startups to redirect valuable time, capital, and talent from innovation to legal compliance. For early-stage firms with limited reserves, this uncertainty actively discourages necessary collaborations and investments.
These all are focused on speculative harms, rather than demonstrated consumer detriment, placing an onerous burden of proof on the innovators, potentially stifling the market’s development in the UK.
In response to this – and to be clear this is not just a UK issue – alternative structures are emerging to mitigate regulatory risk and avoid lengthy merger investigations. But these workarounds are less than optimal for investors and the broader ecosystem as they bypass traditional M&A returns and can erode trust.
Already though, we are seeing the CMA pivoting and doing things differently thanks to a clear steer from the Government. Its 4Ps framework: pace, predictability, proportionality, and process is welcome as is their commitment to building what this means in practice out in the open with support from industry forums such as the Growth and Investment Council (which Startup Coalition sits on).
Clarify Material Influence: A shift in philosophy is needed, moving from presuming harm to demanding concrete, empirical evidence of anti-competitive effects. The CMA should provide clearer “guardrails” on what constitutes “significant other factors” that confer material influence.
The Need for Speed: Lengthy probes, even those concluding in no jurisdiction, have a chilling effect. The welcome introduction of the statutory ‘duty of expedition’ in the Digital Markets Competition and Consumers Act (DMCCA) must be effectively implemented. At the other end – we welcome the newly minted ‘wait and see approach’ but would love to see some of the finer details on what that will look like and where exactly the trigger point for CMA intervention lines.
Narrow Jurisdiction: The broad discretion and questionable application of the ‘share of supply’ test to the digital sphere creates uncertainty. Startups’ investment-seeking strategy of projecting a large Total Addressable Market (TAM) can inadvertently put them in the line of CMA fire. The regulator needs to ensure its market definition doesn’t impede a startup’s ability to innovate and compete.
By creating clarity and predictability in these areas, the CMA can draw a line under a tumultuous period and enable the essential partnerships that will drive the UK’s burgeoning AI ecosystem.