Analyst: Bitcoin holders as a whole have returned to profitability, but it is still too early to confirm the formation of a new bull market.

Baird maintains its "Outperform" rating on AMD, raising the target price from $625 to $1,250, a 100% increase. The adjustment primarily reflects strong growth expectations for AMD’s AI GPU and AI CPU platforms. Baird’s forecast model assumes AMD will capture a 15% share of the AI GPU market, projecting its AI GPU revenue to reach $147 billion by 2030. AMD previously forecast that by 2030, the data center AI accelerator market will hit $1.4 trillion, while the data center CPU market will total $220 billion. AMD’s Helios rack-scale solutions have entered full mass production, and the Helios 500 has been unveiled ahead of schedule. Baird significantly raised AMD’s target price based on its preliminary 2030 earnings per share forecast for the semiconductor company.

According to HTX market data, Bitcoin is currently priced at $65,010. Strategy’s Bitcoin holdings are currently showing an unrealized loss of 13.9%, amounting to approximately $8.8 billion. As of July 20, Strategy holds 843,775 Bitcoin, with a total cost of $63.7 billion and an average cost of $75,476 per Bitcoin.

Prediction market platform Polymarket puts the probability of Bitcoin rising to $70,000 in July at 13%. Additionally, the probability of Bitcoin hitting $75,000 is 1%, while the probability of it falling to $60,000 stands at 14%.

According to tech media Benchlife, NVIDIA (NVDA.O) has issued price hike notifications to its add-in board partners, with the increase covering GDDR6 and GDDR7 memory kits. It is understood that when NVIDIA sells products to these partners, it packages GPU chips and VRAM (video memory) chips into kits for supply, while the partners are responsible for installing the chips onto their self-designed PCBs (printed circuit boards) to produce custom graphics cards. As prices of GDDR6 and GDDR7 memory kits are set to rise again, AIB (Add-in Board, graphics card board) manufacturers are expected to have no choice but to raise graphics card prices accordingly. (Jinshi)

According to official announcements, OKX has collaborated with Elliptic, SlowMist, and OttoSec to release the "H1 2026 Web3 Security and Risk Control Report". The report notes that the focus of Web3 attacks is gradually shifting from smart contract code to more complex scenarios including signature processes, user devices, operational infrastructure, and AI Agents. Data shows that in H1 2026, OKX’s risk control system blocked over 5.7 million high-risk transactions, of which approximately 2.41 million were hack- and theft-related, around 1.48 million were phishing-related, and roughly 990,000 were scam-related. OKX’s Web3 on-chain intelligence label library now holds over 1.1 billion labels, covering more than 420 chains, and has integrated capabilities such as address screening, transaction monitoring, and sanctioned address control into infrastructure including DEX and Exchange OS. Additionally, for user protection, OKX has blocked over 7 million accesses to risky websites, completed more than 200,000 device risk checks, identified over 60,000 high-risk apps, and blocked or alerted on over 4 million high-risk signature operations. The report also introduces "pre-emptive risk control" designs in scenarios such as Exchange OS, Outcomes, RWA, and Agentic Wallet.

According to South Korea’s Electronic News, delivery lead times for major semiconductor manufacturing equipment have recently stretched to twice their original duration. Applied Materials, ASML, Lam Research, Tokyo Electron, and KLA collectively account for roughly 70% of the global semiconductor equipment market, and lead times for their core equipment have generally lengthened to 1.5–2 times the original. Some equipment that normally takes six months to deliver now requires a wait of around one year. Similar delays have hit South Korea’s domestic equipment manufacturers too. One South Korean firm supplying front-end and back-end equipment to TSMC and Micron has extended lead times for some products from three to four months to six to eight months; another back-end equipment supplier to Micron has seen lead times rise to 1.5 times the original, with some units waiting over a year. Equipment delivery delays could impact fabs’ planned launch timelines. Samsung Electronics and SK Hynix have started closely monitoring equipment delivery progress and are considering placing purchase orders earlier than scheduled to reduce the risk of equipment not arriving on time. This current stretch in lead times is distinct from the supply chain disruptions of 2021–2022; it is primarily driven by surging chip demand fueled by AI infrastructure investment, as countries and companies expand fabs simultaneously, triggering a sharp spike in equipment orders. As global fab investment continues, supply shortages related to equipment and fab construction are likely to persist.