Adobe drives AI strategy as Meta builds data centers

Key Takeaways

• AI development is shifting towards efficiency and co-design of hardware and algorithms, exemplified by SkyNet, Medusa (2x LLM speed), and SnapKV (8.2x memory reduction).
• AI could have prevented the alleged $2.7 billion fraud that led to First Brands' bankruptcy in September 2025 by enhancing risk assessment and audit checks.
• CFOs from companies like Adobe, Dataminr, and Huntington Bancshares are integrating AI to speed up decision-making and improve business functions, while also managing data readiness and risk.
• Marc Andreessen believes AI will drive significant economic and job growth, addressing global challenges like declining working-age populations rather than causing widespread job loss.
• The tech hiring market in areas like Utah's Silicon Slopes is experiencing a slowdown, partly attributed to AI and past overhiring, with the national unemployment rate rising to 4.4 percent.
• New "right-to-compute" laws are spreading across US states, supported by figures like President Donald Trump, aiming to limit government regulation on AI and computing, which critics say benefits companies like Meta and Microsoft.
• The UK's AI Security Institute (AISI) tests AI models for safety, but organizations deploying AI retain full liability for compliance, data privacy, and continuous monitoring.
• The S&P 500's target of 7,700 is significantly influenced by artificial intelligence developments and upcoming midterm elections, according to UBS.
• The focus in AI research is moving from simply larger models to more efficient, resource-constrained systems, drawing parallels with the human brain and TinyML concepts.
• European researchers are leveraging AI to address the shortage of skilled hardware engineers, further emphasizing the technology's role in optimizing development processes.

Scarcity Drives True AI Intelligence Not Size

The future of artificial intelligence should focus on efficiency rather than just growing larger. Systems like Voyager 1 and the human brain operate with very limited resources but show great intelligence. Modern AI models, however, use billions of parameters and consume massive amounts of energy, similar to a city. This article argues that scarcity is not a limit but a key factor in developing truly smart and efficient AI systems. Concepts like TinyML and edge inference represent this shift towards getting more from less.

New AI Hardware Designs Boost Efficiency

At the HiPEAC 2026 conference, researchers warned that the current way of building AI models is not sustainable. Chen from the University of Illinois introduced a co-design method to build AI algorithms and hardware together. This approach led to SkyNet, an efficient neural network for drones, and Medusa, which speeds up large language models by two times. Another technique, SnapKV, reduces memory use for conversation history by 8.2 times. European researchers are also using AI to help with the shortage of skilled hardware engineers.

AI Could Have Prevented First Brands 2.7 Billion Dollar Fraud

First Brands, an auto parts company, went bankrupt in September 2025 due to an alleged $2.7 billion fraud. Executives reportedly created fake invoices and hid billions in debt, even selling the same invoice multiple times for much higher amounts. This article explains seven specific ways artificial intelligence could have detected and stopped this fraud. AI tools could assess risks, improve client selection, track cash flow, and automate audit checks to prevent such financial crimes from happening.

Silicon Slopes Tech Hiring Slows But Hope Remains

Hiring in Utah's Silicon Slopes tech hub has slowed down, causing stress for many job seekers like Cody Scott, who searched for over a year. This slowdown is due to factors like AI, past overhiring, higher interest rates, and general economic uncertainty. The national unemployment rate has risen to about 4.4 percent. However, experts from the Utah Department of Workforce Services and Domo see this as a temporary "reset" leading to more sustainable growth in the future.

CFOs Discuss AI Strategy and Scaling Challenges

CFOs from Adobe, Dataminr, and Huntington Bancshares discussed how they are using and scaling AI in their companies at a Fortune webinar on January 27. Tiffany Buchanan of Dataminr sees AI as essential for all business functions. Dan Durn from Adobe focuses on using AI to speed up decision-making and actions. Michael Wasserman of Huntington Bancshares balances fast AI adoption with careful risk management in banking. A key challenge for CFOs is ensuring data is clean and ready for AI models. In 2026, finance teams expect to use AI more for high-level analysis and decision-making.

Marc Andreessen Says AI Boosts Economy Amid Population Decline

Venture capitalist Marc Andreessen believes AI is arriving at the perfect time to address major global issues. He argues that AI will help solve problems like shrinking populations and a lack of new technological innovation over the last 50 years. Andreessen states that AI will not mainly cause job loss but will instead lead to much higher economic and job growth. He highlights that many countries face declining working-age populations, and AI can fill these gaps to maintain productivity.

UK AI Security Institute Helps Buyers But Liability Remains

The UK's AI Security Institute, AISI, helps buyers by testing AI models for safety and security risks. However, the article clarifies that AISI does not take away responsibility from organizations that deploy AI. These organizations must still do their own checks, ensure models work correctly, and follow all rules for data privacy and compliance. AISI acts like a safety certifier, but the user is still responsible for how they use the AI. Buyers should treat AI as high-risk software and continuously monitor models for changes and updates.

Right to Compute Laws Spread Across US States

New "right-to-compute" laws are appearing in US states, starting with Montana in April. These laws aim to make it harder for governments to regulate AI and computing technology. Critics worry these laws mainly help big tech companies like Meta and Microsoft build massive data centers without much oversight. The R Street Institute created model legislation for these laws. While President Donald Trump supports curbing state regulations, nearly 40 other states are also considering laws to limit how businesses use AI.

AI and Elections Influence S&P 500 Target

Jason Katz, a senior portfolio manager at UBS, discussed key factors that could affect the S&P 500's target of 7,700. He explained that artificial intelligence and the upcoming midterm elections are important influences. Katz shared his insights on the "Varney & Co." program.

Sources

• Silicon Darwinism: Why Scarcity Is the Source of True Intelligence
• AI-Hardware Efficiency Gap: Co-Design Solutions from HiPEAC 2026 - News and Statistics
• 7 Ways AI Would Have Stopped First Brands’ $2.7 Billion Fraud
• Months of searching, mounting stress: Why Silicon Slopes hiring feels like it’s at a standstill, and why experts say there’s hope
• What CFOs at Adobe, Dataminr, and Huntington say about scaling AI
• AI Has Arrived At The Perfect Time Amid Shrinking Populations And Low Technological Innovation: Marc Andreessen
• What the UK’s AI Security Institute means for buyers, not builders
• Right-to-Compute Laws Are Spreading Across the US, as Electricity Bills Skyrocket
• AI, midterm elections key factors in S&P 7,700 target: Jason Katz