Headlines this week - Jul 6, 2025

A look at how capital is being deployed across future opportunities

Jul 06, 2025

This week in the future:

1 - Confirmation that the specialized demands of AI are fracturing the cloud market, allowing new players to challenge the established order. Two clear threats for the incumbents, in the news this week:

Oracle catching up: Oracle is rapidly becoming a new hyperscaler by securing massive, multi-billion dollar contracts with AI giants. As detailed in recent reports from the FT, Oracle has landed a landmark $30bn annual deal to provide the computing power for OpenAI's "Stargate" project, a move that will see it build out data centers equivalent to a quarter of the current US capacity.
Coreweave thriving in the "AI hyperscaler" niche: Simultaneously, specialized players like Coreweave are carving out a powerful niche by providing first-to-market access to the most advanced hardware. As reported by CNBC, Coreweave has been the first cloud provider to deploy Nvidia's latest "Blackwell Ultra" (GB300 NVL72) chips, giving it a key differentiator in attracting elite AI labs that need cutting-edge performance.

2 - More support to the idea that AI is placing unprecedented strain on the global electric grid.

Big Tech consumption is growing fast. As detailed in a recent Barclays report, electricity usage at Google and Microsoft soared by over 25% in 2024 for the second consecutive year, driven by AI and cloud computing.
Even nuclear fusion is being considered, to address this. This week we learned about a bet that Google is making on nuclear fusion. They are making a second investment in Commonwealth, a leading contender in the effort to commercialize fusion power, following a 2021 funding round, and they have also committed to buy up to 200MW of power from Commonwealth's planned first commercial plant, expected for early 2030s
Demand volatility is an additional problem. A new, more acute problem is emerging, as highlighted by the CEO of Hitachi Energy in the FT. The training of AI models causes massive, sudden power "spikes" that can be up 10x the normal energy use of a datacenter.This volatility threatens to destabilize global electricity supplies, prompting calls for new regulations to manage this unprecedented and unpredictable demand.

3 - A hyper-competitive "talent war" for the small pool of elite AI researchers has ignited, with compensation packages reaching unprecedented levels:

Meta is the catalyst: As we discussed last week, Meta has launched an aggressive talent acquisition campaign, restructuring its AI unit under a new "Meta Superintelligence Labs" and offering compensation packages reportedly reaching $100m to lure top researchers from rivals like OpenAI and Google.
OpenAI's has reacted by pointing to a "Missionary vs. Mercenary" dilemma: Meta’s moves prompted a visceral reaction from OpenAI's leadership, with one executive stating it felt like "someone has broken into our home." CEO Sam Altman has publicly stated that while OpenAI is "recalibrating comp," and shared his belief that in the long run, "missionaries will beat mercenaries," criticizing Meta's approach as likely to lead to "deep cultural problems."
Skyrocketing Salaries: The bidding war has caused a rapid escalation in wages across the industry, with typical senior AI researcher pay packages in Silicon Valley now ranging from $5m to $7m, a 30% increase since 2022.

4 - As generative AI makes digital content easy to copy, the ultimate competitive differentiator might shift to the physical world, a trend one recent analysis called "The Great Differentiation". With copying now inevitable, the key to sustainable advantage is to create something unique and earned that cannot be easily replicated—emulating the essence of innovation, not just faking the form. This means the most durable moats will be built by integrating AI into tangible, real-world systems and creating unique physical experiences that are true to a company's brand.

5 - The need to catch up in AI challenges Apple’s traditional “walled garden" strategy. After significant internal struggles and delays with its own AI models, Apple would now be in active discussions to license technology from rivals like OpenAI or Anthropic to power the next generation of Siri, according to Bloomberg This potential move represents a major strategic reversal for the company, which has historically prided itself on in-house development. To maintain its privacy-focused brand, Apple has asked these firms to run their models on its own secure "Private Cloud Compute" infrastructure, but the talks have reportedly hit financial hurdles

6 - More news on how AI is revolutionizing drug discovery by automating synthetic proteins design. This week the trend was discussed in a report from The Economist. Rather than merely modifying existing molecules, AI models are now creating entirely novel proteins optimized for specific tasks. This opens the door to a new wave of drugs, such as custom-designed proteins to neutralize snake venom, attack the precursors of Alzheimer's disease, or improve the precision of CRISPR gene-editing tools. This approach is a key driver in compressing R&D timelines by up to 70% and dramatically improving the success rate of drug candidates in early clinical trials.

7 - A recent breakthrough at MIT shows how healthcare is being decentralized by a new generation of cheap, powerful, and increasingly self-managed diagnostic tools. MIT researchers have developed disposable electrochemical sensors, costing only about 50 cents each, that use a DNA-chopping enzyme from the CRISPR system to detect diseases like cancer and HIV from samples like saliva or urine. A key innovation is a polymer coating that gives the sensors a two-month shelf-life without refrigeration, making it possible to deploy them for at-home or point-of-care use in low-resource and remote settings.

8 - Renewable energies are yet another field where China is establishing a formidable geopolitical advantage for the AI era. As detailed in a NYTimes article, China's strategy goes far beyond domestic energy needs. By creating hyper-efficient "cluster manufacturing" hubs where building a battery factory is 6x cheaper than in the US, and by owning the world's 11 largest solar farms, China has built an unassailable lead. It is now leveraging this dominance as a form of "soft power," using over $100bn in foreign clean energy investments since 2023 to forge deep economic and political ties with nations in the Global South, effectively building the energy infrastructure of the future for them and expanding its global influence.

9 - Tesla is struggling: near-term commercial headwinds and an aging product line.

Sales are suffering. As reported by the FT and the WSJ, deliveries have fallen for two consecutive quarters amid rising competition and backlash from Elon Musk's political activism.
However, loyal investors remain committed, betting not on car sales, but on the company's high-risk, high-reward pivot to an AI and robotics powerhouse, with its future valuation hinging on the success of the robotaxi service (where there are also emerging headwinds, e.g. form China) and the Optimus robot.

10 - Nvidia's market dominance is driving a race for alternatives, as highlighted by two developments this week:

Rumors that OpenAI was going to buy Google TPUs (that OpenAI has now denied): OpenAI has publicly denied plans to use Google's in-house AI chips , pushing back against widespread rumors that it was making a major shift away from its reliance on Nvidia hardware. The episode underscores the intense speculation surrounding the strategies of top AI labs as they seek to diversify their supply chains.
Anti-trust regulatory scrutiny reveals SoftBank's ambitions in AI hardware : SoftBank's plan to build an AI infrastructure powerhouse faces a major hurdle, as its $6.5bn acquisition of chip designer Ampere Computing is now under an in-depth investigation by the US Federal Trade Commission (FTC). The probe could delay or even block the deal, which is a cornerstone of SoftBank’s strategy to challenge Nvidia.