Disclosure: To maintain full transparency and independence, we earn commissions on qualifying purchases through affiliate links in this 2026 A2 Hosting Speed Test, ensuring
A2 Hosting‘s consumer hosting business now operates as Hosting.com, so a 2026 buyer may encounter old A2 names, new Hosting.com plans, and “Turbo” terminology in reviews. Benchmark the product currently sold in the cart, not a discontinued plan inferred from an old screenshot.
Current plan context
Hosting.com has advertised entry tiers such as Starter, Plus, Pro, and Max. Recent published allocations have placed Starter around $3.99 per month with one site, 15 GB NVMe, 2 GB RAM, and 2 vCPU; Plus around the same promotional level with two sites, 30 GB, and 3 GB RAM; Pro around $6.99 with 10 sites, 50 GB, 6 GB RAM, and 4 vCPU; and Max around $8.99 with 100 sites, 100 GB, 8 GB RAM, and 4 vCPU. Verify current price, term, and renewal.
Older A2 shared plans used Startup, Drive, Turbo Boost, and Turbo Max. “TurboBoost” in the title should therefore be treated as the LiteSpeed/caching performance concept, not proof that a current plan has identical limits.
Editor's Pick. Our team's current top recommendation for this category. (Affiliate link coming soon — we only link programs we've vetted.)
Test the right stack
Record the checkout product, server region, CPU and RAM allocation, PHP version, WordPress version, web server, cache plugin, CDN state, and test date. If the dashboard exposes Turbo or performance toggles, list them individually.
Install a controlled WordPress fixture: default block theme, optimized hero image, text, navigation, and content cards totaling roughly 1 MB. Remove analytics, ads, remote fonts, security scans, and unrelated cron jobs. Clone the same build to competitors.
Workloads
| Test | Cache state | Load | Why it matters |
|---|---|---|---|
| Homepage repeat | HIT | 1–25 users | LiteSpeed/full-page cache |
| Post repeat | HIT | 1–25 users | Normal readership |
| Purged homepage | MISS | 10 cycles | PHP/database generation |
| Search | BYPASS | 1–10 users | Shared resource allocation |
| WooCommerce cart | BYPASS | 1–10 users | Dynamic session behavior |
Use a staging site and synthetic data. Stop when errors or resource limits appear; the goal is capacity discovery, not denial of service.
Measure TTFB without cherry-picking
Send at least 20 requests per state from a nearby probe. Report median, p75, p95, maximum, and error rate. Repeat from the target audience's region. Say whether connections are reused and which timing fields define TTFB.
For cache HIT tests, verify headers rather than assuming a fast response was cached. For MISS tests, purge before each trial. For BYPASS, use a route or cookie that legitimately excludes caching and ensure no private content is exposed.
A strong cached small-site result generally falls below roughly 200–400 ms near-region, while uncached WordPress can be higher. The stable p95 under expected concurrency is more important than a record-setting minimum. These thresholds are evaluation criteria, not claimed Hosting.com numbers.
Load-test procedure
Ramp from one virtual user through five, 10, and 25, with realistic wait time. Validate response bodies and record HTTP 429, 5xx, and timeout responses. Monitor CPU, RAM, entry processes, PHP workers, database activity, and disk I/O if exposed.
Page caching can flatten the cached curve because PHP is skipped. Search and cart requests reveal the actual account resources. A Pro or Max result cannot be attributed to Starter merely because both use NVMe and LiteSpeed.
Run each stage long enough to reach a stable pattern. Brief bursts can fit inside queues or boost behavior and overstate sustainable throughput. Publish the duration and request count with any result.
What “Turbo” can and cannot do
LiteSpeed page caching can serve completed pages efficiently. NVMe storage can reduce storage latency. OPcache avoids recompiling PHP, and an object cache can reduce database queries when available. These layers help different paths.
They cannot repair a slow external API, a page builder shipping megabytes of JavaScript, an unindexed query, or a plugin that executes on every request. A “20x faster” style marketing statement is not a substitute for a controlled workload and should never be converted into an expected load time.
Use one compatible page-cache plugin. Configure exclusions for carts and personalized pages, then test purging after edits. Combining LiteSpeed Cache with another full-page cache often creates conflicts.
Browser performance
Run WebPageTest or Lighthouse from a fixed location with a fixed mobile connection. Use at least three runs and compare medians. Inspect LCP, CLS, total blocking time, transfer size, and the waterfall.
If TTFB is good and LCP is slow, optimize the largest image and critical CSS. If browser main-thread time is high, reduce JavaScript. If MISS TTFB is slow while HIT is strong, investigate PHP, database queries, and plan limits.
Hosting.com versus competitors
Hostinger also uses LiteSpeed and usually offers a low long-term price, but hPanel differs from cPanel-style workflows. SiteGround uses NGINX Dynamic Cache with documented HIT/MISS/BYPASS headers and a high renewal. Bluehost emphasizes guided setup, automatic caching, Redis on eligible WordPress products, and Cloudflare.
Hosting.com's published resource allocations are a welcome aid if they remain visible at purchase. The rebrand is the weakness: older reviews and plan names create confusion. Support should confirm which former A2 features exist on the selected 2026 product.
Pros and cons
Pros
- Current plans advertise concrete RAM, vCPU, NVMe, and site allocations.
- LiteSpeed-oriented caching can serve anonymous WordPress efficiently.
- Higher tiers provide substantial resources for shared-style hosting.
- Familiar WordPress and hosting tools remain available.
Cons
- A2 and Hosting.com naming makes old comparisons unreliable.
- Promotional and renewal terms require current verification.
- Cache-hit marketing can conceal dynamic limits.
- Large site allowances still share each plan's resources.
FAQ
Is A2 Hosting gone?
The hosting business is presented under Hosting.com, while older A2 branding and reviews remain online. Use the current checkout name and plan specifications for purchasing decisions.
Is TurboBoost faster than ordinary shared hosting?
LiteSpeed and full-page caching can be much faster for repeat anonymous requests. Dynamic performance still depends on CPU, RAM, PHP, database work, and plugins.
Which plan should a store use?
Choose a higher-resource plan with daily backups and staging, then test cart and checkout. Entry tiers are better for cached blogs than concurrent dynamic sessions.
Can benchmark results survive the rebrand?
Only if the server stack and resource allocation are unchanged and disclosed. A result tied to an old plan should not be assigned to a new product without verification.
Verdict
Hosting.com retains the speed-oriented appeal associated with A2, and its current published resource figures make plan-level testing possible. The honest test separates LiteSpeed HIT performance from dynamic BYPASS capacity and names the exact product. Buy only after confirming renewal and current features, then benchmark the real account during its refund window rather than relying on an old TurboBoost chart.