May saw the Q1 earnings season continue, with many of our portfolio companies reporting. We continue to see numbers and management commentary that suggest that AI demand still far outstrips supply. Outside of the results, we also saw a number of compute deals that reflect the ongoing shortage \u2013 most notably when Anthropic signed a deal with xAI\u2019s Colossus 1 data center. We\u2019ve been writing for the past few months about how compute-constrained Anthropic and OpenAI are \u2013 deals like this are clear evidence of that, with no sign that we\u2019re anywhere close to over-building. Google announced its latest token metrics (a very good indicator for us of underlying AI demand) \u2013 it\u2019s now processing 3.2 quadrillion tokens per month, up from 480 trillion tokens per month a year ago and 9.7 trillion two years ago.<\/p>\n\n\n\n
We are starting to see companies\u2019 own supply chain capabilities play a bigger role in results \u2013 with component shortages limiting the revenue some companies can achieve. Arista Networks (owned) was an example of a company that was clearly constrained by supply, although we still expect revenue and profit upgrades throughout the year. Nvidia (owned) also reported results in May, which largely delivered what we have come to expect: solid outperform and uplift, with particular strength in compute and networking. Taking a step back, the numbers remain quite extraordinary \u2013 quarterly revenue over $1.4T80 billion and growth of $851T3T year-on-year \u2013 with no sign yet of the law of large numbers. In terms of supply chain capabilities, Nvidia\u2019s execution and ecosystem positioning remain exceptional and that remains a significant competitive advantage. Nvidia increased its supply commitments by another $1.4T24 billion this quarter to $1.4T119 billion.<\/p>\n\n\n\n
These commitments include TSMC, memory vendors, network vendors, and component manufacturers. That\u2019s an extraordinary number that few companies can match. The reality is that one of Nvidia\u2019s most important near-term competitive advantages is simply that they have secured more supply than anyone else. We also saw AMD\u2019s (owned) results clearly show that they are benefiting from agent-based AI and increased demand for CPUs. After providing a long-term market forecast back in November, AMD doubled it down \u2013 and now expects a CAGR of 35% through 2030 \u2013 with demand for agent-based CPUs as the clear driver of the shift. Just like Intel\u2019s forecast a few weeks ago, AMD\u2019s management commented on the potential CPU\/GPU ratio moving towards 1:1.<\/p>\n\n\n\n
Outside of AI, end-market data remains mixed. US auto volumes remain weak, and while we haven't seen it in the data yet, oil prices are clearly expected to result in some consumer weakness, and we remain generally cautious around companies exposed to the auto and PC\/smartphone markets.<\/p>\n\n\n\n
The fund increased by 2.2% in May and 8% so far this year.<\/p>\n\n\n\n